SYNTHETIC APERTURE RADAR FOR ESTIMATING THE ABOVE-GROUND BIOMASS OF A SPARSE SAVANNA WOODLAND

Savannas are important ecosystems in the global carbon cycle due to their high net primary productivity. SAR interferometry and backscatter are investigated for estimating biomass in a savanna woodland in Belize, Central America. Single-pass InSAR data used are C-band (AIRSAR) and X-band (Intermap); SAR backscatter data used are fully polarimetric L- and P-band (AIRSAR). Results show that high P- and L-band backscatter is not only observed for high biomass vegetation, but also for leafy palmetto. Both C-band and X-band InSAR show a clear trend in vegetation patterns although both underestimate vegetation heights due to the heterogeneity of the vegetation cover. The scattering phase centre is lower for the X-band than for C-band over areas of sparse woodland.

The implications of these findings are that, despite the proven capability of SAR to estimate biomass in denser forests, this capability does not extend to low density woodlands. Savannas cover 20% of the earth's land surface in the tropics, and are predominently found in developing countries. These results suggest that remote sensing of biomass of savannas using SAR could result in a significant underestimation of above-ground biomass in savanna areas and hence an under estimation of their carbon sequestration potential.